Statistical analyses
Statistical analyses were carried out via R software (version 3.4.2). Diversity indices (i.e., Shannon-Wiener index and Richness) and Plelou’s evenness of functional genes and OTUs were calculated. Two-tailed Student’st -tests on diversity indices, relative abundances of each functional gene and gene category, and relative abundance of each OTU and higher taxonomic levels were performed between warmed and control samples. P -values were adjusted by the Benjamini & Hochberg (BH) correction with the false discovery rate of 0.10. The multi-response permutation procedure (MRPP) and analysis of similarity (ANOSIM) were performed via mrpp and anosim functions in veganpackage in R to examine overall differences in functional gene composition or OTU composition between warmed and control samples. If significant differences were found, then the similarity percentage (SIMPER) analysis, via the simper function in the veganpackage in R, was applied to reveal which functional genes or OTUs caused the differences as well as their contributions. The principal coordinate analysis (PCoA) was performed to visualize the overall composition of functional genes and OTUs. The canonical correlation analysis (CCA) and the variation partitioning analysis (VPA), via the cca function in vegan package in R, were performed to reveal which environmental factors were closely linked to differences between control and warmed samples and their contributions. We only selected environmental factors with the variance inflation factor (VIF) < 20 in the CCA model and tested the overall significance of the CCA model by ANOVA. Relationships betweenR eco and relative abundances of fungal functional genes for C degradation were examined by linear regression.